Research map / Label-efficient and robust learning / Noisy labels

Robust Classification via Regression for Learning with Noisy Labels

ICLR 2024 · Label-efficient and robust learning › Noisy labels · approach: Sample selection and co-training

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes reformulating classification as regression with a bounded-output loss (and label correction) to make deep networks robust to noisy labels, combining loss reweighting and label correction.

Problem

Deep neural networks overfit to label noise in large-scale datasets, reducing generalization in learning with noisy labels.

Approach

Recasts classification as a regression problem so that loss reweighting and label correction are unified, aiming to be robust to noisy labels; the abstract is truncated so the exact mechanism is only partially stated.

Closest papers by meaning

• Mitigating Label Noise through Data Ambiguation (AAAI 2024 · similarity 0.91)
• Two Wrongs Don’t Make a Right: Combating Confirmation Bias in Learning with Label Noise (AAAI 2023 · similarity 0.91)
• Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noise (NeurIPS 2018 · similarity 0.90)
• Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization (NeurIPS 2025 · similarity 0.90)
• Robust Loss Functions under Label Noise for Deep Neural Networks (AAAI 2017 · similarity 0.90)
• Optimal Transport Reweighting for Robust Learning under Spurious Correlations and Label Noise (NeurIPS 2026 · similarity 0.89)